physics benchmark
PIQA Leaderboard
PIQA is a benchmark dataset for physical commonsense reasoning in natural language, with 21,000 multiple choice questions about everyday physical interactions.
Updated Sep 5, 2026
Models11
Model coverage11
MetricScore
EvidenceB
PIQA Ranking
Higher score ranks better on this benchmark.
Sort by
Rank | Model | Score | Percentile | Participants | Evidence | Evaluated |
|---|
| Rank01 | ModelMI | Score88.60% | Percentile100.00% | Participants11 | EvidenceC | Evaluated |
| Rank02 | ModelNR | Score84.44% | Percentile90.00% | Participants11 | EvidenceC | Evaluated |
| Rank03 | ModelGO | Score83.20% | Percentile80.00% | Participants11 | EvidenceC | Evaluated |
| Rank04 | ModelGO | Score81.70% | Percentile70.00% | Participants11 | EvidenceC | Evaluated |
| Rank05 | ModelGO | Score81.00% | Percentile60.00% | Participants11 | EvidenceC | Evaluated |
| Rank06 | ModelGO | Score81.00% | Percentile50.00% | Participants11 | EvidenceC | Evaluated |
| Rank07 | ModelMI | Score81.00% | Percentile40.00% | Participants11 | EvidenceC | Evaluated |
| Rank08 | ModelGO | Score78.90% | Percentile30.00% | Participants11 | EvidenceC | Evaluated |
| Rank09 | ModelGO | Score78.90% | Percentile20.00% | Participants11 | EvidenceC | Evaluated |
| Rank10 | ModelMI | Score77.60% | Percentile10.00% | Participants11 | EvidenceC | Evaluated |
| Rank11 | ModelBA | Score55.20% | Percentile0.00% | Participants11 | EvidenceC | Evaluated |
PIQA Highlights
The leading models and scores on this benchmark.
PIQA Score Distribution
A closer view of the leading scores on this benchmark.
The Top AI Models for PIQA
The first five results on this benchmark, with official price and output speed added where the model identity can be matched.
What is PIQA?
What PIQA measures and how its scores work.
PIQA (Physical Interaction: Question Answering) is a benchmark dataset in the physics category that tests AI systems with multiple choice questions requiring physical world knowledge.
It measures the ability to choose the most appropriate solution for physical interactions in everyday situations, with performance reported as Score in ratio units.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- PIQA
- Modality
- text
- Primary category
- physics
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- overall
LLM Stats. Benchmark scores retain their original unit. Overall score eligibility is shown separately.
FAQ
Common questions about PIQA.
Which model scores highest on PIQA?
Phi-3.5-MoE-instruct is currently ranked first with 88.60%.
What are the top three models on PIQA?
The current leaders are Phi-3.5-MoE-instruct (88.60%), Hermes 3 70B (84.44%), and Gemma 2 27B (83.20%).
Which PIQA model has the lowest official input price?
Phi 4 Mini has the lowest matched official input price at $0.08 input / $0.30 output per 1M tokens.
Which models are fastest among PIQA results?
The fastest matched records are Phi-3.5-mini-instruct (23.00 tok/s via Azure).
Does the highest score result prove overall model quality?
No. This benchmark measures one defined capability or task. The overall LLMBoard score uses a separate aggregation across eligible benchmark evidence.